Using the workloads added in PERF-8816, measure the foreground workload throughput impact of primary-driven index builds on DSC across a range of scan-phase and bulk-load-phase write throttle settings, and recommend defaults for both parameters.

This is the ticket that closes Goal 1 of the scope document, which requires that an index build be no more impactful to foreground operation throughput on DSC than the same index build on ASC.

Deliverables

• A curve of foreground throughput impact vs. index build duration for each phase's throttle, measured on DSC, with ASC as the baseline.
• Recommended default values for both parameters, or a recommendation that they ship disabled by default and be used as an operational lever only.
• Confirmation of whether the ASC-parity bar in Goal 1 is met.

Note that throttling index build writes trades Goal 1 against Goal 2 (index build duration within 1.5x of the same build on ASC). The analysis needs to state explicitly where the chosen defaults sit on that tradeoff.